To: djgpp AT delorie DOT com
Subject: Re: inlined functions with inline assembly using -O3
Message-ID: <19961208.085246.6727.2.aclemmer@juno.com>
References: <199612080035 DOT SAA17374 AT ghgcorp DOT ghgcorp DOT com>
From: aclemmer AT juno DOT com (Aaron m Clemmer)
Date: Sat, 07 Dec 1996 20:54:39 EST

On Sat, 7 Dec 1996 17:09:46 +0000 "Bachtel" <rbachtel AT ghgcorp DOT com>
writes:

>Anyway, the problem is that you're using abosolute instead of 
>relative jumps. How do you use relative jumps? Perceptive question! 
>Try something like:
	Will do... Hopefully, this will work... hey, something weird
happened the other day... I have yet another function that is declared
inline.  Anyway, I started wondering if inline functions are actually
inlined with optimizations off (I could probably find that out easily
enough <g>), so I went ahead and pasted the code directly into the
function I was timing.  What suprised me was that the block of code that
was being time was actually _slower_ when I pasted the code directly in,
as opposed to letting GCC do it by itself!  Maybe oneday I'll go compare
the assembly output for both ways... =)

thanks,

aaron